Fuel 86 (2007) 829–834

www.fuelfirst.com

Dewatering of fine coal slurries by selective heating with microwaves

a,*
M.S. Seehra , A. Kalra b, A. Manivannan a

a
Department of Physics, West Virginia University, Morgantown, WV 26506-6315, United States
b
Department of Mechanical & Aerospace Engineering, West Virginia University, Morgantown, WV 26506-6315, United States

Received 6 February 2006; received in revised form 13 June 2006; accepted 1 August 2006
Available online 12 September 2006

Abstract

Comparative experimental studies on the dewatering of a fine coal slurry sample containing 52% moisture by microwave heating
vis-à-vis thermal heating are reported. For thermal heating, thermogravimetric analysis (TGA) using different masses and different heat-
ing rates were employed. Similar experiments were carried out for dewatering by microwave heating using a conventional oven oper-
ating at 2.45 GHz and 800 W power. These quantitative experiments show that microwave heating is an order of magnitude more
efficient than thermal heating, although about 10% of the residual moisture, perhaps trapped in the capillaries of the coal particles,
is not removed even by microwave heating. Based on these results, a bench scale dewatering unit was designed and tested in which
coal slurry in kg quantities was fed on a conveyor belt. The efficiency of this unit for dewatering fine coal slurries was estimated
to be over 80% of the theoretical limit. A cost estimate of about $3/ton for the power consumed for 10% reduction of moisture is
made.
 2006 Elsevier Ltd. All rights reserved.

Keywords: Dewatering; Microwaves; Fine coal slurries

1. Introduction loss of valuable carbons. Consequently, a need exists to find

an effective process for dewatering fine coal slurries.
The presence of a variety of impurity phases (clays, In this work, we present the results of our studies on the
quartz, pyrite, etc.) in mined coals requires extensive grind- dewatering of a fine coal slurry using conventional micro-
ing and preparation to separate the useful carbonaceous waves at 2.45 GHz. The basis for microwave drying is the
material from the impurities [1]. In coal preparation, water selective absorption of microwave power by water mole-
is used extensively which must be removed as much as pos- cules since the carboneaus matter is known to absorb
sible before coal can be shipped and burnt in power plants. microwave energy relatively poorly [3]. This process has
Mechanical dewatering methods (e.g. centrifuges) are gen- been discussed in earlier studies by Lyons et al. [4], Lind-
erally ineffective for finer coal particles [2]. Thermal dewa- roth [5], Chatterjee and Misra [6], Perkin [7] and Standish
tering methods are not cost effective since they heat both et al. [8] in connection with drying of coals and other por-
water and coals, the latter resulting in the loss of BTU value ous materials and in recent studies by Marland et al. [9] in
of coals. Because of these difficulties, most of the finer coal connection with coal grindability and by Lester and King-
fractions are abandoned and relegated to settling ponds as man [10]. The power P absorbed by a material from elec-
slurries. These settling ponds represent serious environmen- tromagnetic (em) radiation of frequency f can be written
tal threat for the surrounding communities in addition to as [3–11]

2
P ¼ 55:63  1012 jE0 j f e00 ; W=m3 : ð1Þ
*
Corresponding author. Tel.: +1 304 293 3422x1473; fax: +1 304 293
5732. Here E0 is the amplitude of em radiation and e00 is the loss
E-mail address: [email protected] (M.S. Seehra). factor of the material at frequency f. At f = 2.45 GHz,

0016-2361/$ - see front matter  2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.fuel.2006.08.015
830 M.S. Seehra et al. / Fuel 86 (2007) 829–834

e00 = 10 for water compared to e00 = 0.1 for coal, making P 3. Results and discussion
about 100 times more for water than that for coal [3]. This
provides the basis for the selective absorption of em radia- 3.1. Thermal heating
tion by water present in coal, without significantly affecting
the temperature of the carbon particles. In addition, micro- The results of TGA studies in terms of % weight loss
waves penetrate deeply in the coal carbons thus heating against temperature using samples of masses of 50, 100
coals inside out as compared to conventional thermal heat- and 200 mg and heating rate of 3 C/min are shown in
ing from the outside inwards using the standard heat trans- Fig. 1. These experiments were done in a continuous man-
fer mechanisms. The results presented below compare ner without taking the samples out of the balance and the
conventional thermal heating with microwave heating of data points at steps of about 3 C are shown only for visual
a fine coal slurry sample and show that energy efficiency clarity although in the actual experiments, the data are
above 80% of the theoretical limit is obtained by micro- taken every 3 s. It is evident that larger mass requires
wave dewatering. Results of our experimental simulations higher temperature for the same mass loss, as expected.
of this process for possible industrial applications are also The total weight loss of nearly 45% due to decrease in
described. the moisture is slightly less than the moisture content
of the sample obtained by proximate analysis (Table 1),
suggesting that some of the moisture, most likely in the
2. Pertinent experimental details
capillaries, has not been removed [9,10].
The effect of different heating rates on the elimination of
The sample of fine coal slurry used in the experiments
moisture is also evident from the DSC (differential scanning
here was obtained from Consol Energy Inc. (Pittsburgh).
calorimetry) studies at the heating rates of 5 and 10 C/min
This represents a discarded sample of a bituminous coal
(Fig. 2) showing that higher thermal heating rates require
sent to a settling pond as a refuse from a coal preparation
higher temperatures for the same level of moisture reduc-
plant. The slurry sample was allowed to settle in the con-
tion (as in Fig. 1). The data of Fig. 1 was used to determine
tainer and excess water at the top was skimmed off. The
the activation energy Ea associated with the moisture reduc-
proximate analysis of the remaining settled sample (carried
tion using the Flynn–Wall method [12–14] for a constant
out by Galbraith Laboratories, Knoxville, TN) given in
Table 1, shows nearly 53% moisture and about 27%
carbon, the remaining being ash. Scanning electron micros- 100 Coal Sent to Ponds
copy of the sample gave an average particle size to be about -Different Weight

280 lm (micrometer) with most of the particles being in the 90

100 mg
size range of 120–440 lm. Thermogravimetric analysis
Weight (%)

80
(TGA) was done by a Mettler 3000 microbalance using mil- 200 mg

ligram samples. For microwave dewatering, a commercial 70 Heat rate

3 O C/min
oven (GE Model JES838SH001) with maximum power
50 mg
output of 800 W was used. To measure the power used in 60
an experiment, a power meter was inserted in the circuit
50
between the power outlet and the microwave oven. The 40 60 80 100 120
slurry sample was placed on a Styrofoam plate, heated Temperature (°C)
for a given time, then taken out of the oven and weighed.
Fig. 1. Remaining weight (%) from the TGA experiments against
The sample is then reinserted in the oven and the pro-
temperature for three different masses using heating rate of 3 C/min.
cess repeated till constant weight representing nearly com- The lines connecting the data points are for visual aid.
plete loss of moisture was achieved. Details of the bench
scale unit for simulating a commercial system are given
later. 0
-10 a
Heat flow

-20
Table 1
ο
Proximate analysis of settled coal slurry sample -30 a- 5 C/min b
ο
Sample ID Analysis As received Dry basis -40 b- 10 C/min
Samples sent to ponds Loss on drying 52.53 % -50
Carbon 26.59 % 56.01 %
Hydrogen 6.50 % 1.31 % -60
Nitrogen <0.5 % <0.5 %
Sulfur 1.42 % 2.99 % 40 60 80 100 120
ο
Chlorine 230 ppm 485 ppm Temperature ( C)
Ash 15.231 % 32.08 %
Fig. 2. Results of differential scanning calorimetry for two different
Oxygen (by diff) 50.24 % 7.56 %
heating rates.
 M.S. Seehra et al. / Fuel 86 (2007) 829–834 831

weight loss for the different heating rates of b = 3, 10 and 100

Power Consumption (kWh)

50 gm 0.14
20 C/min. Using the temperatures T (K) for the constant 100 gm
150 gm 0.12
weight (say 80%), a plot of ln b vs. 1/T yields a st. line 90 200 gm
(not shown) with slope Ea/R, where R = 8.314 J/K mol is 0.10

% Weight
the gas constant. The analysis carried out for the constant 0.08
80
weights at 80%, 70% and 60% yields Ea  9 kcal/mol in 0.06
agreement with the expected value of 9.7 kcal/mol for the 0.04
specific heat of vaporization of water at 100 C. 70
0.02
0.00
3.2. Microwave heating 60
0 2 4 6 8
Using the microwave oven described earlier and initial Time (min)
sample weights of 50, 100, 150 and 200 g, the % weight Fig. 4. Remaining weight (%) vs. time using microwave heating for four
changes as a function of time are shown in Fig. 3. As noted different initial masses of the slurry sample. The right ordinate shows that
earlier, in these experiments the sample, spread out on a the power consumed varies linearly with time, the larger masses requiring
Styrofoam plate, was heated for 10 s interval, taken out more time and more power for complete dewatering. The data points were
taken for heating in 30 s intervals.
of the oven and quickly weighed on a digital balance and
reinserted in the oven for the next reading. The power con-
sumed in the process is also plotted in Fig. 3 which clearly The efficiencies of microwave and thermal heating were
increases linearly with time. A similar experiment was compared in the following way using the same amount
repeated for 30 s heating intervals and results are shown (20 g) of the slurry sample. For thermal heating, the sample
in Fig. 4. It is evident that the larger heating time reduces was placed on a glass dish and kept in an oven set at 110 C
the time needed for complete elimination of the moisture (with an air flow to take away eliminated moisture) and
by nearly a factor of two. This may be mostly due to the taken out every 30 s interval for weighing. The microwave
expected higher temperatures achieved for 30 s heating. heating process was repeated similarly as described earlier
Additional caution is warranted since the samples lost using the same amount of initial mass of 20 g. The compar-
some heat during the weighing process (lasting about 5 s) ative weight losses as a function of time (Fig. 5) clearly
outside the oven. However, absence of significant scatter show the significant advantages of microwave heating in
in the data from smooth variations in Figs. 3 and 4 pro- terms of reduced time by a factor of nearly ten.
vides confidence in the data. The primary conclusion from From a practical point of view, it is important to know
Figs. 3 and 4 is that the power consumed is proportional to how microwave radiation affects the useful properties of
the time microwaves are on irrespective of the mass of the coals. This issue has been investigated by Lester and King-
sample wherein it takes more time to dewater a sample with man [10] using microwave power up to 8 kW at 2.45 GHz
larger mass vis-à-vis a smaller mass. Another conclusion is for exposure times of several seconds. This study showed
that the maximum weight loss of around 40% is smaller that ‘‘the proximate composition, rank and intrinsic reac-
than the moisture content of about 50% of the sample tivity did not change significantly’’. In our studies, we car-
(Table 1) meaning that moisture in the capillaries is most ried out X-ray diffraction studies (CuKa radiation with
likely not removed. This is in agreement with the conclu- k = 0.1542 nm) of the ‘‘before’’ and ‘‘after’’ sample of
sions from earlier studies [10]. microwave heating (Fig. 6). The sharp lines are due to var-
ious minerals whereas the broad background peak centered
around 2h  25 is due to the carbonaceous component of
100 50 gm 0.18
Power Consumed (KWh)

100 gm
150 gm 0.16
90 100
200 gm 0.14
% Weight

0.12
80 0.10 90
0.08
% Weight

Thermal
70 0.06 80
0.04
0.02 70
60 Microwave
0.00
0 2 4 6 8 10 12 14 60
Time (min)

Fig. 3. Remaining weight (%) vs. time using microwave heating for four 0 10 20 30 40 50 60 70 80
different initial masses of the slurry sample. The right ordinate shows that Time (min)
the power consumed varies linearly with time, the larger masses requiring
more time and more power for complete dewatering. The data points were Fig. 5. Comparison of the efficiencies of thermal heating vis-à-vis
taken for heating in 10 s intervals. microwave heating using 20 g sample in each case. For details, see text.
832 M.S. Seehra et al. / Fuel 86 (2007) 829–834

Intensity (arbitary unit) Coal sent to Ponds 0.16

0.14

Sample Before MW Treatment

0.12

Power (kWh)
0.10 10 sec
interval
0.08
0.06 30 sec
MW Treated Samples 0.04 interval
0.02
0.00
0 0 20 40 60 80 100 120
10 20 30 40 50 60 70 80
O
Moisture removed (g)
2-Theta( )
Fig. 7. Using the data of Figs. 3 and 4 till a constant weight is achieved,
Fig. 6. Room temperature X-ray diffractrograms for the slurry sample power consumed is plotted against moisture removed in grams for 10 s
‘‘before’’ and ‘‘after’’ dewatering by microwave heating. (Fig. 3) and 30 s (Fig. 4) heating durations. The dotted lines represent the
‘‘eyeball’’ slopes dP/dm.

the slurry [11]. No significant change such as presence of

any new lines in the ‘‘after’’ sample are indicated although only 54% (0.75/1.4) for the 10 s data. The improved effi-
admittedly X-ray diffraction is not adequate to determine ciency observed for longer heating times is discussed in
changes in the various carbonaceous component of the connection with the results obtained below with the
coals. bench-scale unit.
For practical reasons, the amount of power consumed
for dewatering is an important issue. Using the data of 3.3. Bench-scale microwave dewatering unit
Figs. 3 and 4, the power P consumed vs. moisture removed
m for both the 10 s and 30 s heating is shown in Fig. 7. The To simulate a commercial microwave dewatering unit,
dotted lines are drawn to estimate the slope dP/dm. The we next designed a conveyor belt system employing the
small non-zero intercept extrapolated to m = 0 may repre- same microwave oven (Fig. 8), with the following modifica-
sent the power lost to the sample holder and surroundings. tions. First, the two opposite ends of the oven were opened
The slope dP/dm = 1.33 · 103 kW h/g (9.8 · 104 kW h/ to feed the coal slurry on a conveyor belt so that kg quan-
g) for 10 s (30 s) heating. Assuming the same dewatering tities can now be fed. The block diagram of the unit is
rate for larger masses of wet coal, 1 kW h should vaporize shown in Fig. 9. A commercially available conveyor belt
0.75 kg (1.02 kg) of water for 10 s (30 s) heating. Theoreti- (Taconic TFE glass) is fed by two rollers driven by a
cally, energy H needed to convert m grams of water at motor. To protect the operator from microwave radiation,
25 C to steam at 100 C is given by the complete unit is protected by a commercial microwave
shield (Diamond Manufacturing) and a circuit was
H ¼ mfð100  25Þð4:18Þ þ 2256g J ¼ 2:569m kJ: ð2Þ
designed to operate the system remotely. A portable micro-
For m = 1 kg, Eq. (2) gives H = 2.569 · 106 J = 0.71 kW h. wave detector (MD 2000) was used to check for any leaks
Conversely, 1 kW h should vaporize 1.40 kg of water, the of the microwaves outside the apparatus.
maximum theoretical value. Thus we get 73% (1.02/1.4) For testing this system, a measured amount (kg quan-
efficiency of dewatering from the 30 s data of Fig. 7 and tity) of the coal slurry sample is spread on the conveyor

Fig. 8. (a) Picture of the bench-scale microwave dewatering units with microwave shield in place. (b) Close up of the unit without the microwave shield
showing the conveyor belt for feeding wet coal into the microwave oven.
 M.S. Seehra et al. / Fuel 86 (2007) 829–834 833

Conveyor
Belt Wet Dry
Roller Coal Coal

Microwave
Oven

Motor
Power
Supply for Dry
Motor Coal

Fig. 9. Block diagram of the bench-scale microwave dewatering unit.

belt. Experiments were done for different speeds of the belt tude (0.8 kW) is slightly above the mean value obtained by
(thus varying the resident time of the sample inside the Lindroth but lower than the maximum obtained in his
oven) and power consumed and moisture lost measured. experiments at higher power levels. So our results at the
In Fig. 10, results of these experiments are shown. Clearly lower power levels are similar and provide an independent
as the belt speed is lowered (increasing the resident time), verification of the high levels of efficiency obtainable in
the % moisture lost increases along with the power con- microwave drying of a fine coal slurry which contains finer
sumed. The slope of these plots is the smallest (most effi- particles but higher levels of moisture than coals.
cient parameters) for the belt speed around 13 cm/min. An approximate cost estimate for dewatering by 10% a
Using the mass of the sample, this belt speed and % mois- ton of fine coal slurry yields about $3/ton assuming 83%
ture lost corresponds to 1.16 kg/kW h of moisture lost rep- efficiency and power cost of 0.04$/kW h. This represents
resenting efficiency of about 83% (1.16/1.40). For other belt only the cost of the power used and does not include cap-
speeds, the efficiency is lower than 83%. This increase in ital and labor costs in such an operation.
efficiency from the earlier 73% efficiency is probably due
to the fact that larger quantities of samples are used thus 4. Summary
reducing the power lost to the surroundings. In the earlier
studies of Lindroth on the microwave drying of various fine The laboratory scale and bench scale studies reported
coals [5], he used higher power magnetrons (9.7 kW) at here on dewatering of a fine coal slurry containing about
2.45 GHz to simulate an industrial drying process and 52% moisture using conventional microwave radiation at
obtained efficiencies in the range of 52–97% with a mean 2.45 GHz and 800 W have shown efficiencies 80% com-
efficiency of 77%. This large variation in efficiency resulted pared to theoretical maximum. About 10% of the moisture
from changes in process parameters such as coal type, feed appear to remain perhaps in the capillaries of the fine coal
rate, belt speed, initial moisture level and the like. The best particles. It is argued that overall characteristics of micro-
efficiency of about 83% obtained in our experiments on one wave dewatering represent significant advantages over
coal slurry using power levels lower by an order of magni- dewatering by thermal heating. Using the bench-scale unit
tested in the experiments with the coal slurry fed on a con-
4 veyer belt, it is demonstrated that a system can be designed
Belt speed (cm/min)

Belt Speed 6
Power consumed 8
10
and operated without radiation hazard to the operator.
12
14
The adoption of this technology for industrial applications
16
18
however depends significantly on the economics of the coal
0.05 20 market, a subject beyond the scope of this work.
Power (kWh)

0.04
0.03 5. Disclaimer
0.02
0.01
This report was prepared as an account of work spon-
30 40 50 60 70 80 sored by an agency of the United States Government. Nei-
% Moisture lost ther the United States Government nor any agency thereof,
Fig. 10. Power consumed and belt speed (cm/min) vs. % moisture lost. nor any of their employees, make any warranty, express
The region of the least slope (13 cm/min) yields the optimum speed (see or implied, nor assume any legal liability or responsibil-
text). ity for the accuracy, completeness, or usefulness of any
834 M.S. Seehra et al. / Fuel 86 (2007) 829–834

information, apparatus, product, or process disclosed, or [2] Dzinomwa GPT, Wood CJ, Hill DJT. Fine coal dewatering using pH
represents that its use would not infringe privately owned and temperature-sensitive super-absorbent polymers. Polym Adv
Technol 1997;8:767–72.
rights. Reference herein to any specific commercial prod- [3] Marland S, Merchant A, Rowson N. Dielectric properties of coal.
uct, process, or service by trade name, trademark, manu- Fuel 2001;80:1839–49.
facturer, or otherwise does not necessarily constitute or [4] Lyons DW, Hatcher JD, Sunderland JE. Drying of a porous medium
imply endorsement, recommendation, or favoring by the with internal heat generation. Int J Heat Mass Transfer 1972;15:
United States Government or any agency thereof. The 897–905.
[5] Lindroth DP. Microwave drying of fine coal. US Bureau of Mines
views and opinions of authors expressed herein do not nec- Report # 9005, US Department of Interior, 1985.
essarily state or reflect those of the United States Govern- [6] Chatterjee I, Misra M. Electromagnetic and thermal modeling
ment or agency thereof. of microwave drying of fine coal. Miner Metall Process 1991;8:110–
4.
Acknowledgements [7] Perkin RM. The heat and mass transfer characteristics of boiling
point drying using radio frequency and microwave electromagnetic
fields. Int J Heat Mass Transfer 1980;23:687–95.
The authors thank the following persons for assistance [8] Standish N, Worner H, Kaul H. Microwave drying of brown coal
with this project: Phillip Tucker for designing and building agglomerates. J Microwave Power EM Energy 1988;23:171–5.
the electrical circuit for remote control operation of the [9] Marland S, Han B, Merchant A, Rowson N. The effect of microwave
bench scale unit, Steve Pachol for providing the coal slurry radiation on coal grindability. Fuel 2000;79:1283–8.
[10] Lester E, Kingman S. The effect of microwave preheating on five
sample and Sukanya Ranganathan for the SEM studies. different coals. Fuel 2004;83:1941–7 and references therein.
Financial support for this work was provided by the Center [11] Babu VS, Farinash L, Seehra MS. Carbon in diesel particulate
for Advanced Separation Technologies, established by a matter: structure, microwave absorption and oxidation. J Mater Res
contract from the US Department of Energy (Contract # 1995;10:1075–8.
DE-FC26-02NT4160T). [12] Flynn JH, Wall LA. A quick, direct method for the determination of
activation energy from thermogravimetric data. Polym Lett
1966;4:323–8.
References [13] Ibrahim MM, Hopkins E, Seehra MS. Thermal and catalytic
degradation of commingled plastics. Fuel Process Technol
[1] Wheelock TD, Markuszewski R. Coal preparation and cleaning. In: 1996;49:65–73.
Cooper BR, Ellingson WA, editors. The science and technology of [14] Ismail MKI, Rogers SL. Comparisons between fullerene and forms of
coal and coal utilization. New York: Plenum Press; 1984. p. 47–123. well-known carbons. Carbon 1992;30:229–39.